[0001] The present invention relates generally to a wireless communication system, and in
particular, to a transmission/reception apparatus using transmission antenna diversity
to compensate for degradation due to fading.
[0002] In a wireless communication system, time and frequency diversity is one effective
techniques for suppressing fading. Among known techniques for antenna diversity, a
space-time block code proposed by Vahid Tarokh extends transmission antenna diversity
proposed by S.M. Alamouti so that two or more antennas can be used. The proposal made
by Tarokh is disclosed in a paper "Space Time Block Coding From Orthogonal Design,"
IEEE Trans. on Info., Theory, Vol. 45, pp. 1456-1467, July 1999, and the proposal
made by Alamouti is disclosed in a paper "A Simple Transmitter Diversity Scheme For
Wireless Communications," IEEE Journal on Selected Area in Communications, Vol. 16,
pp. 1451-1458, Oct. 1998.
[0003] Fig. 1 is a block diagram illustrating a structure of a transmitter using a space-time
block code according to the prior art. The transmitter is proposed by Tarokh, and
as illustrated, is comprised of a serial-to-parallel (S/P) converter 110 and an encoder
120. In this structure, the transmitter uses three antennas 130, 132 and 134.
[0004] Referring to Fig. 1, the S/P converter 110 groups 4 input symbols into one block,
and provides the block to the encoder 120. The encoder 120 makes 8 combinations with
the 4 symbols, and delivers the 8 combinations to the 3 transmission antennas 130,
132 and 134 for 8 time intervals. The 8 combinations can be expressed in an 8×3 encoding
matrix which is defined as

where g
3 represents an encoding matrix of symbols transmitted via 3 transmission antennas,
and s
1, s
2, s
3 and s
4 represent 4 input symbols to be transmitted.
[0005] The encoder 120 applies negative and conjugate to 4 input symbols, and outputs the
result values to the 3 antennas 130, 132 and 134 for 8 time intervals. In this case,
symbol sequences output to the antennas, i.e., rows, are orthogonal with one another.
[0006] More specifically, in a first time interval, 3 symbols s
1, s
2, and s
3 in a first row are delivered to the 3 antennas 130, 132 and 134, respectively. Likewise,
in the last time interval, 3 symbols s
4*, s
3* and s
2* in the last row are delivered to the 3 antennas 130, 132 and 134, respectively. That
is, the encoder 120 sequentially delivers symbols in an M
th row of the encoding matrix to an M
th antenna.
[0007] Fig. 2 is a block diagram illustrating a structure of a receiver for receiving a
signal transmitted from the transmitter of Fig. 1. As illustrated, the receiver is
comprised of a plurality of reception antennas 140 and 142, a channel estimator 150,
a multi-channel symbol arranger 160, and a detector 170.
[0008] Referring to Fig. 2, the channel estimator 150 estimates channel coefficients representing
a channel gain from transmission antennas to reception antennas, and the multi-channel
symbol arranger 160 collects reception symbols from the reception antennas 140 and
142, and provides the collected reception symbols to the detector 170. The detector
170 then calculates a decision statistic for all possible symbols with hypotheses
symbols determined by multiplying the reception symbols by the channel coefficients,
and detects transmission symbols by threshold detection.
[0009] The space-time block coding technique proposed by Alamouti, though complex symbols
are transmitted through 2 transmission antennas, obtains a diversity order equivalent
to the number of transmission antennas, i.e., the maximum diversity order, without
inflicting a loss on a rate. The devices of Figs. 1 and 2 proposed by Tarokh by extending
this technique obtain a maximum diversity order using a space-time block code in the
form of a matrix having orthogonal rows. However, since the devices transmit 4 complex
symbols for 8 time intervals, they suffer a loss of the rate by 1/2. In addition,
since 8 time intervals are required in completely transmitting one block (having 4
symbols), reception performance become poor due to a variation in a channel environment
within a block in the case of fast fading.
[0010] When complex symbols are transmitted using 3 or more antennas as mentioned above,
2N time intervals are required in order to transmit N symbols, resulting in a loss
of a rate. The loss of a rate also causes an increase in latency.
[0011] It is, therefore, an object of the present invention to provide a transmission/reception
apparatus for securing a maximum diversity order and a maximum rate without a loss
of a rate in a wireless communication system using 3 transmission antennas.
[0012] According to a first aspect of the present invention, there is provided a transmitter
for transmitting complex symbols in a wireless communication system. The transmitter
comprises three transmission antennas; and an encoder for grouping 4 input symbols
into 4 combinations each including three symbols so that the 4 input symbols are transmitted
only once at each antenna and each time interval, and delivering the 4 combinations
to the three transmission antennas for 4 time intervals; wherein two or more symbols
selected from the 4 input symbols are phase-rotated by predetermined phase values,
respectively.
[0013] According to a second aspect of the present invention, there is provided a receiver
for receiving complex symbols in a wireless communication system. The receiver comprises
a symbol arranger for receiving signals received via at least one reception antenna
from thee transmission antennas, for four time intervals; a channel estimator for
estimating three channel gains representing channel gains from the three transmission
antennas to the reception antenna; first and second decoders for calculating metric
values for all possible sub-combinations each including two symbols by using the channel
gains and the signals received by the symbol arranger, and detecting two symbols having
a minimum metric value; and a parallel-to-serial converter for sequentially arranging
two symbols detected by the first and second decoders.
[0014] According to a third aspect of the present invention, there is provided a transmitter
for transmitting complex symbols in a wireless communication system. The transmitter
comprises three transmission antennas; and an encoder for grouping 3 input symbols
into 3 combinations each including three symbols so that the 3 input symbols are transmitted
only once at each antenna and each time interval, and delivering the 3 combinations
to the three transmission antennas for 3 time intervals; wherein two or more symbols
selected from the 3 input symbols are phase-rotated by predetermined phase values,
respectively.
[0015] According to a fourth aspect of the present invention, there is provided a receiver
for receiving complex symbols in a wireless communication system. The receiver comprises
a symbol arranger for receiving signals received via at least one reception antenna
from three transmission antennas, for three time intervals; a channel estimator for
estimating three channel gains representing channel gains from the three transmission
antennas to the reception antenna; and a decoder for calculating metric values for
all possible symbol combinations each including three symbols by using the channel
gains and the signals received by the symbol arranger, and detecting three symbols
having a minimum metric value.
[0016] The above and other objects, features and advantages of the present invention will
become more apparent from the following detailed description when taken in conjunction
with the accompanying drawings in which:
Fig. 1 is a block diagram illustrating a structure of a transmitter using a space-time
block code according to the prior art;
Fig. 2 is a block diagram illustrating a structure of a receiver for receiving a signal
transmitted from the transmitter of Fig. 1;
Fig. 3 is a block diagram illustrating a structure of a transmitter using a space-time
block code according to a first embodiment of the present invention;
Fig. 4 is a block diagram illustrating a structure of a receiver for receiving a signal
transmitted by the transmitter of Fig. 3;
Fig. 5 illustrates a simulation result showing a variation in a minimum coding gain
for 2 phase values when QPSK is used in the first embodiment of the present invention;
Fig. 6 illustrates a QPSK constellation which is phase-rotated by 45°;
Fig. 7 is a graph illustrating a comparison between the block coding technique according
to the first embodiment of the present invention and the conventional techniques in
terms of a bit error rate (BER) for a signal-to-noise ratio (SNR);
Fig. 8 is a block diagram illustrating a structure of a transmitter using a space-time
block code according to a second embodiment of the present invention;
Fig. 9 is a block diagram illustrating a structure of a receiver for receiving a signal
transmitted by the transmitter of Fig. 8;
Fig. 10 illustrates a simulation result showing a variation in a minimum coding gain
for 2 phase values when QPSK is used in the second embodiment of the present invention;
and
Fig. 11 is a graph illustrating a comparison between the block coding technique according
to the second embodiment of the present invention and the conventional techniques
in terms of a bit error rate (BER) for a signal-to-noise ratio (SNR).
[0017] Several preferred embodiments of the present invention will now be described in detail
with reference to the annexed drawings. In the following description, a detailed description
of known functions and configurations incorporated herein has been omitted for conciseness.
In addition, the terms used in the following description are defined considering their
functions in the invention. Therefore, a definition of the terms must be given based
on the overall contents of the specification.
[0018] The invention phase-rotates a part of a complex transmission signal to secure a maximum
diversity order and a maximum rate, and introduces a partial orthogonal structure
to simplify a decoding scheme. In particular, the invention proposes two embodiments
for an optimal block code which is available when 3 transmission antennas are used.
A first embodiment is to optimize a diversity order and a rate, and the second embodiment
is to minimize transmission latency. The two embodiments will be separately described
below. Although a structure and operation of phase-rotating two transmission symbols
will be described below, two or more transmission symbols can be phase-rotated to
accomplish the invention.
First Embodiment
[0019] In the first embodiment of the invention, 4 input symbols are transmitted via 3 antennas
for 4 time intervals, and this can be expressed in an encoding matrix defined as

[0020] As is well known, a receiver using ML (Maximum Likelihood) decoding employs a scheme
for calculating a metric value with a reception signal for all possible symbols based
on a channel gain from a transmission antenna to a reception antenna, and detecting
a symbol that minimizes the calculated metric value.
[0021] In a receiver receiving the symbols of Equation (2), if a channel gain from an i
th transmission antenna to one reception antenna is defined as h
i, a metric value corresponding to a particular symbol combination c
t is expressed as

where r
t represents a signal received in a t
th time interval, and c
t represents a particular symbol combination created in a t
th time interval. When the encoding matrix of Equation (2) is applied to Equation (3),
the receiver determines a symbol combination that minimizes Equation (4) below, for
all possible symbol combinations.

where r
1, r
2, r
3 and r
4 are signals received at the receiver for 4 time intervals, respectively, and h
1, h
2 and h
3 are channel gains representing channel coefficients from 3 transmission antennas
to a reception antenna.
[0022] In order to simplify an ML detection scheme of a receiver, as many crossover terms
as possible must be removed from Equation (4) so that symbol sequences, i.e., rows,
transmitted via transmission antennas are orthogonal with one another. For that purpose,
only crossover terms are enumerated below.

[0023] It is well known by Tarokh that when 4 symbols are transmitted using a 4×3 encoding
matrix, all crossover terms appearing during ML detection can be removed. However,
it is possible to allow at least first and third antennas h
1 and h
3 to have orthogonality by removing at least 2 terms, i.e., C
1 and C
2, from Equation (5).
[0024] In order to secure a maximum diversity order, 4 transmission symbols must appear
only once at each antenna and each time interval, and shown in Equation (6) are 4
examples of 4×3 encoding matrixes satisfying such a condition. Other encoding matrixes
can be formed by mutually permuting rows or columns of the 4 matrixes.

[0025] Shown in Equation (7) below is an example of an encoding matrix to which negative
and conjugate are applied in order to eliminate 2 crossover terms, i.e., C
1 and C
2, of Equation (5) for the encoding matrixes of Equation (6).

[0026] Shown in Equation (8) below are possible examples of an encoding matrix in which
rows are partially orthogonal while securing a maximum diversity order.

where x
1, x
2, x
3 and x
4 are arbitrarily arranged after negative and conjugate are applied to s
1, s
2, s
3 and s
4. Specifically, Equation (7) shows a second matrix of Equation (8) in which x
1=s
1, x
2=s
2, x
3=-s
4*, x
4=-s
3*·
[0027] When at least 2 crossover terms C
1 and C
2 are removed using the encoding matrixes of Equation (8), an ML detection scheme of
a receiver can be simplified even further. For example, if Equation (4) is expressed
again by applying the encoding matrix of Equation (7), minimizing Equation (4) is
identical to minimizing Equation (9) and Equation (10) below. This is possible because
a metric of Equation (9) and a metric of Equation (10) are independent of each other.


where "Min(a,b)(y(a,b))" means determining "a,b" that minimizes "y(a,b)," and "Re{}"
means calculating a real component for a complex number in braces. In addition, C
1 and C
2 become 0 as mentioned above, and C
3=h
3*h
2-h
3h
2* and C
4=h
3h
2*-h
3*h
2=-C
3. Moreover, R
1=r
1h
1*+r
2*h
2+r
3*h
3, R
2=r
1h
2*-r
2h
1+r
4h
3*, R
3=r
2*h
3+r
4h
1*-r
3*h
2, and R
4=r
1h
3*-r
3*h
1-r
4h
2*.
[0028] Using Equation (9) and Equation (10), a receiver decouples a part for decoding a
pair of s
1 and s
3 according to Equation (9) from a part for decoding a pair of s
2 and s
4 according to Equation (10), thereby further simplifying its structure.
[0029] Meanwhile, when input symbols were generated by BPSK (Binary Phase Shift Keying),
the above-stated encoding matrix always has a diversity order of 3. However, when
a symbol mapping scheme of a 3
rd or higher order using a complex constellation, i.e., QPSK (Quadrature Phase Shift
Keying), 8PSK (8-ary Phase Shift Keying) and 16PSK (16-ary PSK), is used, transmission
symbols become complex symbols, so a diversity order is reduced to 2. Therefore, the
invention secures a maximum diversity order 3 by phase-rotating each of 2 symbols
that determine different metric values, among 4 symbols, by a predetermined phase
value. Then, symbols finally transmitted via 3 antennas are expressed as

[0030] Equation (11) shows an encoding matrix for phase-rotating s
1 and s
4 among input symbols s
1, s
2, s
3 and s
4 of Equation (7) by θ
1 and θ
2, respectively. In another case, it is possible to rotate a symbol pair of (s
1,s
2), (s
3,s
4) or (s
2,s
3) related to different matrixes. Although phase values by which the 2 symbols are
rotated respectively are different from or identical to each other, a diversity order
is always maintained at 3. Likewise, if 2 symbols that determine different metric
values are phase-rotated by a predetermined phase value even for the other encoding
matrixes of Equation (8), final encoding matrixes can be obtained.
[0031] A transmitter and a receiver using the encoding matrixes described above are illustrated
in Figs. 3 and 4, respectively.
[0032] Fig. 3 is a block diagram illustrating a structure of a transmitter using a space-time
block code according to a first embodiment of the present invention. As illustrated,
the transmitter is comprised of a serial-to-parallel (S/P) converter 210, phase rotators
220 and 222, an encoder 230, and three transmission antennas 240, 242 and 244.
[0033] Referring to Fig. 3, the S/P converter 210 groups 4 input symbols s
1, s
2, s
3 and s
4 into one block, and provides the block to the encoder 230. Two symbols s
1 and s
4 selected from the block are rotated by predetermined phase values θ
1 and θ
2, respectively, by the phase rotators 220 and 222 before being provided to the encoder
230. The 2 symbols are selected so that they are related to different metrics at a
receiver. The encoder 230 makes 4 combinations each including 3 symbols, with symbols
of one block including the 2 phase-rotated symbols, and delivers the 4 combinations
to the 3 transmission antennas 240, 242 and 244 for 4 time intervals.
[0034] In order to obtain a maximum diversity order, the encoder 230 makes the combinations
so that the 4 input complex symbols should be transmitted only once at each antenna
and each time interval. In addition, the encoder 230 makes the combinations by applying
negative and conjugate to the input symbols so that symbol sequences delivered to
each antenna should be orthogonal with one another. The reason for phase-rotating
2 symbols selected from the input symbols is to obtain a maximum diversity order even
when the input symbols are complex symbols.
[0035] If the 4 combinations transmitted via the 3 antennas are expressed in a 4×3 matrix,
symbols in an M
th row of an encoding matrix are sequentially delivered to an M
th antenna. That is, in an n
th time interval, symbols in an n
th column are simultaneously delivered to the 3 antennas.
[0036] For example, when s
1 and s
4 among 4 input symbols s
1, s
2, s
3 and s
4 are phase-rotated by θ
1 and θ
2, respectively, an output of the encoder 230 can be expressed in a 4×3 encoding matrix
of Equation (11) above. When the encoding matrix of Equation (11) is used, 3 symbols
ejθ1s1, s
2 and
ejθ4s4 in a first row are delivered to the 3 antennas 240, 242 and 244, respectively, in
a first time interval and symbols s
3,
ejθ4s4 and s
2 in the last 4
th row are delivered to the 3 antennas 240, 242 and 244, respectively, in the last 4
th time interval.
[0037] A transmitter for transmitting the matrix of Equation (11) has been described so
far. However, in a modified embodiment of the present invention, a transmitter may
multiply the matrix of Equation (11) by a unitary matrix before transmission.
[0038] Fig. 4 is a block diagram illustrating a structure of a receiver for receiving a
signal transmitted by the transmitter of Fig. 3. The receiver according to a first
embodiment of the present invention includes two ML decoders 340 and 345, which operate
independently.
[0039] Referring to Fig. 4, a channel estimator 320 estimates channel coefficients, i.e.,
channel gains h
1, h
2 and h
3, from the 3 transmission antennas 240, 242 and 244 to reception antennas 310 and
315, and a symbol arranger 330 collects signals r
1, r
2, r
3 and r
4 received via each of the reception antennas 310 and 315 for 4 time intervals.
[0040] If the number of reception antennas is 1, the symbol arranger 330 collects signals
r
1, r
2, r
3 and r
4 received for 4 time intervals. This is because the transmitter transmitted symbols
of one block for 4 time intervals. When two or more reception antennas are used, the
symbol arranger 330 forms a matrix by collecting received signals. In this case, the
symbol arranger 330 arranges signals received via one reception antenna in one row,
and arranges signals received via another reception antenna in another row. Although
the receiver has herein multiple reception antennas 310 and 315, a description of
the invention will be made with reference to a case where one reception antenna is
used, for simplicity.
[0041] When it is desired to restore 4 symbols s
1, s
2, s
3 and s
4 transmitted from a transmitter, the first decoder 340 detects s
1 and s
3 according to the channel gains and the reception signals, and the second decoder
345 detects s
2 and s
4 in the same manner. In this way, the 4 symbols are simultaneously detected by the
first and second decoders 340 and 345. The detected symbols are represented by s'
in order to distinguish them from their original symbols.
[0042] An operation the first decoder 340 will now be described in a case where the encoding
matrix of Equation (11) is used. In the first decoder 340, a symbol generator 350
generates all possible sub-combinations s
1 and s
3, and a phase rotator 360 phase-rotates one, s
1, of the generated symbols by the same phase value θ
1 as that used by a transmitter, and outputs
ejθ1s1. A metric calculator 370 determines metric values by calculating Equation (9) for
all symbol sub-combinations including one phase-rotated symbol with the estimated
channel gains h
1, h
2 and h
3 and the reception signals r
1, r
2, r
3 and r
4. A minimum metric detector 380 then detects s
1' and s
3' having minimum metric values among the metric values.
[0043] Such an operation is performed in the same manner in the second decoder 345. When
the first decoder 340 detects s
1' and s
3' and the second decoder 345 detects s
2' and s
4' in this manner, a parallel-to-serial (P/S) converter 390 sequentially arranges the
detected symbols and outputs a symbol combination of s
1', s
2', s
3' and s
4'.
[0044] A phase value used to phase-rotate symbols in the transmitter and the receiver of
Figs. 3 and 4 must be determined according to a minimum coding gain of error matrixes.
The "error matrix" is a matrix in which differences between symbols containing errors
detected at a receiver and originally transmitted symbols are arranged, and the "minimum
coding gain" means the product of all eigen values of an error matrix.
[0045] Fig. 5 illustrates a simulation result showing a variation in a minimum coding gain
for 2 phase values when QPSK is used in the first embodiment of the present invention.
In Fig. 5, an x-axis and a y-axis represent 2 phase values, respectively, and a z-axis
represents a minimum coding gain of an error matrix. If a phase value becomes a multiple
of 90°, a minimum coding gain becomes 0. This is because if a QPSK constellation is
rotated by 90°, it becomes its original constellation.
[0046] It can be understood from the result of Fig. 5 that when all phase values exist at
around 45°, the minimum coding gain becomes flat. Therefore, a phase value preferable
in the first embodiment of the invention is 45°. Fig. 6 illustrates a QPSK constellation
which is phase-rotated by 45°. As illustrated, the phase-rotated symbols are situated
on a real axis or an imaginary axis. According to the first embodiment of the invention,
a preferable phase rotation range is between 21° and 69° for QPSK, between 21° and
24° for 8PSK, and is 11.25° for 16PSK, centering on 45°. However, the invention is
not restricted to the figures, and the preferable phase rotation range shall be set
according to characteristics of the system.
[0047] Fig. 7 is a graph illustrating a comparison between the block coding technique according
to the first embodiment of the present invention and the conventional techniques in
terms of a bit error rate (BER) for a signal-to-noise ratio (SNR). In Fig. 7, a curve
410 shows efficiency in the case where an 8×3 encoding matrix with orthogonal rows
proposed by Tarokh is used, a curve 420 shows efficiency in the case where 2 antennas
are used as proposed by Alamouti, a curve 430 show efficiency in the case where a
4x3 encoding matrix having a phase value optimized according to the first embodiment
is used, and a curve 440 shows efficiency in the case where a 4×3 encoding matrix
having a non-optimized phase value is used. As illustrated, a block code having a
phase value optimized according to the first embodiment has a lower BER in the same
SNR environment.
Second Embodiment
[0048] In the second embodiment of the invention, 3 input symbols are transmitted via 3
antennas for 3 time intervals. Compared with the first embodiment, the second embodiment
further decreases transmission latency.
[0049] As mentioned above, in order to secure a maximum diversity order, each symbol must
appear only once in each time interval of each antenna, and a unique 3×3 encoding
matrix satisfying such a condition is given by

[0050] An error matrix of a space-time block code using the encoding matrix of Equation
(12) can be expressed as

where C
33 is a transmission encoding matrix, and E
33 is a matrix representing determined symbols containing errors. In Equation (13),
a coding gain of D
33 is 3d
1d
2d
3-d
13-d
23-d
33. Thus, if d
1=d
2=d
3 or d
2=0 and d
1=-d
3, then the coding gain becomes 0. In this case, a diversity order is lower than 3
which is the number of the transmission antennas, thus incurring a large loss in performance.
[0051] In the second embodiment of the invention, in order to prevent a coding gain from
becoming 0, two symbols selected from three symbols are rotated by a predetermined
phase value, and this can be expressed in an encoding matrix defined as

[0052] Herein, s
1 and s
2 among 3 input symbols s
1, s
2 and s
3 are phase-rotated by -θ
1 and -θ
2, respectively. Then, a coding gain of a space-time block code using the encoding
matrix of Equation (14) always becomes 3.
[0053] If a metric value is calculated with channel gains h
1, h
2 and h
3 from 3 transmission antennas to a reception antenna for Equation (14), it becomes

[0054] A receiver then determines symbols s
1 to s
3 that minimize Equation (15).
[0055] Fig. 8 is a block diagram illustrating a structure of a transmitter using a space-time
block code according to a second embodiment of the present invention. As illustrated,
the receiver is comprised of an S/P converter 510, two phase rotators 520 and 522,
an encoder 530, and three transmission antennas 540, 542 and 544.
[0056] Referring to Fig. 8, the S/P converter 510 groups 3 input symbols s
1, s
2 and s
3 into one block, and provides the block to the encoder 530. Two symbols s
1 and s
3 in the block are rotated by predetermined phase values -θ
1 and -θ
2, respectively, by the phase rotators 520 and 522 before being provided to the encoder
530. The encoder 530 makes 3 combinations each including 3 symbols, with symbols of
one input block, and delivers the 3 combinations to the 3 transmission antennas 540,
542 and 544 for 3 time intervals.
[0057] In other words, the encoder 530 applies negative and conjugate to 3 input complex
symbols, and outputs the result values for 3 time intervals. Herein, the encoder 530
sequentially delivers symbols in an M
th row of an encoding matrix to an M
th antenna. That is, the encoder 530 simultaneously delivers symbols in an n
th column in an n
th time interval.
[0058] For example, when s
1 and s
3 among 3 input symbols s
1, s
2 and s
3 are phase-rotated by -θ
1 and -θ
2, respectively, an output of the encoder 530 can be expressed in a 3×3 encoding matrix
of Equation (14) above. When the encoding matrix of Equation (14) is used, 3 symbols
e-jθ1s1,
e-jθ2s2 and
s3 in a first row are delivered to the 3 antennas 540, 542 and 544, respectively, in
a first time interval, and symbols e
-jθ2s2 , s
3, and
e-jθ1s1 in the last 3
rd row are delivered to the 3 antennas 540, 542 and 544, respectively, in the last 3
rd time interval.
[0059] Fig. 9 is a block diagram illustrating a structure of a receiver for receiving a
signal transmitted by the transmitter of Fig. 8. Although the receiver has herein
multiple reception antennas 610 and 615, a description of the invention will be made
with reference to a case where one reception antenna is used, for simplicity.
[0060] Referring to Fig. 9, a channel estimator 620 estimates channel coefficients, i.e.,
channel gains, from the 3 transmission antennas 540, 542 and 544 to reception antennas
610 and 615, and a multi-channel symbol arranger 630 collects signals received via
the reception antennas 610 and 615. If the number of reception antennas is 1, the
symbol arranger 630 forms one block by collecting signals received for 3 time intervals.
When two or more reception antennas are used, the symbol arranger 630 forms a matrix
by collecting signals received via the reception antennas for 3 time intervals. In
this case, the symbol arranger 630 arranges signals received via one reception antenna
in one row, and arranges signals received via another reception antenna in another
row.
[0061] An ML decoder 640 then restores 3 desired symbols every third time intervals with
the channel gains from the channel estimator 620 and the reception signals from the
symbol arranger 630. The ML decoder 640 is comprised of a symbol generator 650, phase
rotators 660 and 665, a metric calculator 670, and a minimum metric detector 680.
[0062] An operation the ML decoder 640 will now be described in a case where the encoding
matrix of Equation (14) is used. The symbol generator 650 generates all possible combinations
s
1, s
2 and s
3, and outputs them one by one in each time interval, and the phase rotators 660 and
665 phase-rotate two symbols s
1 and s
2 selected from the symbols output from the symbol generator 650 by the same phase
values -θ
1 and -θ
2 as those used by the transmitter, respectively, and output
e-jθ1s1 and
e-jθ2s2. A combination of
e-jθ1s1 and e
-jθ2s2 will be called a symbol combination.
[0063] The metric calculator 670 determines metric values for all symbol combinations by
multiplying the channel gains h
1, h
2 and h
3 by all the possible symbol combinations generated by the symbol generator 650 according
to a predetermined method and using reception signals r
1, r
2 and r
3 arranged by the symbol arrangers 660 and 665. An operation of the metric calculator
670 is performed in accordance with Equation (15). The minimum metric detector 680
then detects symbol combinations s
1', s
2' and s
3' having minimum metric values among the metric values.
[0064] A coding gain of the encoding matrix shown in Equation (14) depends upon a phase
value used in phase-rotating symbols. Fig. 10 illustrates a simulation result showing
a variation in a minimum coding gain for 2 phase values when QPSK is used in the second
embodiment of the present invention. In Fig. 10, an x-axis and a y-axis represent
2 phase values, respectively, and a z-axis represents a minimum coding gain of an
error matrix. A minimum coding gain 0 means a loss of a diversity gain.
[0065] It can be understood from the result of Fig. 10 that when phase values are a multiple
of 30°, the minimum coding gain becomes 0. Therefore, two phase values are determined
so that the minimum coding gain is maximized. As a difference between the two phase
values becomes greater, their performances are superior. Preferable phase values according
to the second embodiment of the invention become a multiple of 30°, such as 30°, 60°,
90°, 120°, 150° and 180°.
[0066] Fig. 11 is a graph illustrating a comparison between the block coding technique according
to the second embodiment of the present invention and the conventional techniques
in terms of a bit error rate (BER) for a signal-to-noise ratio (SNR). In Fig. 11,
a curve 710 shows efficiency in the case where a non-phase-rotated encoding matrix
is used, a curve 720 shows efficiency in the case where a 3×3 encoding matrix having
a phase value (7.48°) optimized according to the second embodiment is used, and a
curve 730 show efficiency in the case where a 3×3 encoding matrix having a non-optimized
phase value (24° and 25°) is used. As illustrated, a block code having a phase value
optimized according to the second embodiment has a lower BER in the same SNR environment.
[0067] As described above, the invention can obtain a maximum diversity order without a
loss of a rate, and is robust against fast fading by decreasing transmission latency.
In particular, the first embodiment of the invention can simplify a decoding scheme
by allowing some rows of an encoding matrix to become orthogonal with each other,
and the second embodiment of the invention can further reduce transmission latency.
[0068] While the invention has been shown and described with reference to a certain preferred
embodiment thereof, it will be understood by those skilled in the art that various
changes in form and details may be made therein without departing from the scope of
the invention as defined by the appended claims.
1. Transmitter for transmitting complex symbols in a wireless communication system, comprising:
three transmission antennas; and
an encoder for grouping N input symbols into N combinations each including three symbols
so that the N input symbols are transmitted only once from each antenna and at each
time interval, and delivering the N combinations to the three transmission antennas
for N time intervals;
wherein at least two re symbols selected from the N input symbols are phase-rotated
by predetermined phase values.
2. Transmitter according to claim 1, wherein N is 4.
3. Transmitter according to claim 2, wherein the number of the selected symbols is 2,
and the selected symbols are related to different metrics during decoding at a receiver.
4. Transmitter according to claim 3, wherein for quadrature phase shift keying (QPSK),
the phase values range from 21° to 69°, centering on 45°.
5. Transmitter according to claim 3, wherein for 8-ary phase shift keying (8PSK), the
phase values range from 21° to 24°.
6. Transmitter according to claim 3, wherein for 16-ary phase shift keying (16PSK), the
phase values are 11.25°.
7. Transmitter according to claim 2, wherein the encoder produces four combinations by
applying negative and conjugate to four symbols so that two symbol sequences among
three symbol sequences delivered to each antenna for four time intervals are orthogonal
with each other.
8. Transmitter according to claim 7, wherein the four combinations are each comprised
of the four input symbols and constitute matrixes each having four rows and three
columns, as follows

where x
1, x
2, x
3 and x
4 are four input symbols including two phase-rotated symbols.
9. Transmitter according to claim 1, wherein N is 3.
10. Transmitter according to claim 9, wherein three combinations are each comprised of
three input symbols and constitute a matrix having three rows and three columns, as
follows

where s
1, s
2 and s
3 are the three input symbols, and θ
1 and θ
2 are phase values of s
1 and s
2, respectively.
11. Transmitter according to claim 9, wherein the phase values are a multiple of 30°,
and are determined so that a difference between the phase values becomes maximized.
12. Receiver for receiving complex symbols in a wireless communication system, comprising:
a symbol arranger for receiving signals received via at least one reception antenna
from three transmission antennas, for four time intervals;
a channel estimator for estimating three channel gains representing channel gains
from the three transmission antennas to the reception antenna;
first and second decoders for calculating metric values for all possible sub-combinations
each including two symbols by using the channel gains and the signals received by
the symbol arranger, and detecting two symbols having a minimum metric value; and
a parallel-to-serial converter for sequentially arranging two symbols detected by
the first and second decoders.
13. Receiver according to claim 12, wherein the first and second decoders each comprise:
a symbol generator for generating all possible sub-combinations each including two
symbols;
a phase rotator for phase-rotating one symbol selected from the two symbols by a predetermined
phase value;
a metric calculator for calculating a metric value for symbol sub-combinations including
the phase-rotated symbol with the signals received by the symbol arranger and the
channel gains; and
a detector for detecting two symbols having a minimum metric value by using the calculated
metric values.
14. Receiver according to claim 13, wherein the first decoder detects two symbols for
minimizing a metric value calculated by

where s
1 and s
3 are two symbols to be detected, θ
1 is a phase value of s
1, R
1=r
1h
1*+r
2*h
2+r
3*h
3, R
3=r
2*h
3+r
4h
1*-r
3*h
2, C
3=h
3*h
2-h
3h
2*, h
1, h
2 and h
3 are channel gains estimated for three transmission antennas, and r
1, r
2, r
3 and r
4 are signals received for four time intervals.
15. Receiver according to claim 13, wherein the second decoder detects two symbols for
minimizing a metric value calculated by

where s
2 and s
4 are two symbols to be detected, θ
2 is a phase value of s
2, R
2=r
1h
2*-r
2h
1+r
4h
3*, R
4=r
1h
3*-r
3h
1-r
4h
2*, C
4=h
3h
2*-h
3*h
2, h
1, h
2 and h
3 are channel gains estimated for three transmission antennas, and r
1, r
2, r
3 and r
4 are signals received for four time intervals.
16. Receiver for receiving complex symbols in a wireless communication system, comprising:
a symbol arranger for receiving signals received via at least one reception antenna
from three transmission antennas, for three time intervals;
a channel estimator for estimating three channel gains representing channel gains
from the three transmission antennas to the reception antenna; and
a decoder for calculating metric values for all possible symbol combinations each
including three symbols by using the channel gains and the signals received by the
symbol arranger, and detecting three symbols having a minimum metric value.
17. Receiver according to claim 16, wherein the decoder comprises:
a symbol generator for generating all possible symbol combinations each including
three symbols;
two phase rotators for phase-rotating two symbols selected from the three symbols
by predetermined phase values;
a metric calculator for calculating metric values for symbol combinations including
the phase-rotated symbols with the signals received by the symbol arranger and the
channel gains; and
a detector for detecting three symbols having a minimum metric value by using the
calculated metric value.
18. Receiver according to claim 17, wherein the decoder detects three symbols for minimizing
a metric value calculated by

where s
1, s
2 and s
3 are three symbols constituting a symbol combination, θ
1 and θ
2 are phase values of s
1 and s
2, respectively, h
1, h
2 and h
3 are channel gains for three transmission antennas, and r
1, r
2 and r
3 are signals received for three time intervals.
19. Transmitter for transmitting complex symbols in a wireless communication system, comprising:
M transmission antennas; and
an encoder for grouping N input symbols into N combinations each including M symbols
so that the N input symbols are transmitted only once from each antenna and at each
time interval, and delivering the N combinations to the M transmission antennas for
N time intervals;
wherein at least two symbols selected from the N input symbols are phase-rotated
by predetermined phase values.